470 



"tXl/l ^ ^ ■ 



< thus somewhat difTerent lo wlial would 



tbilow from Table II of the jdevious paper. 



Tlie deviations are gra|)l)icall_y represented 



in fifj,'. 1. It is very strikinj-'. liow much the 



accuracy of the measurements lias been 



increased since previous delerniinalions. The 



circumstances under which they were carried 



! out (very constant room-tempeiature, entire 



absence, of leakages, etc.) were extremely 



favourable and, as great care was bestowed 



, on the measurements, it appears that they 



luue reached the full measure of accuracy 



of which they are capable. 



Tiie accuracy of 50V0 to ,0^00. which 

 ' would follow from the excellent agreement 

 of our results with the formula can only 

 ^ be a relative accuracy in view of the un- 

 ;E cerlainty of the normal volume (see above). 

 It must therefore be ascribed to an accidental 

 concurrence of favourable circumstances, that 

 the agreement with Schat.kwijk's observa- 

 tions is so very close. As the matter stands, 

 the portion of the isothermal determined by 

 him is continued without an^y discontinuity 

 ' by that of our experiments. 



The figure also contains the value of 

 /»;,! for 20"^ C. and 100 atmospheres, which 

 J would follow from Am.ag.^t's observations 

 according to the principles developed in 

 Comm. 71 and which has been calculated by 

 "°i a formula given by Schalkwijk. The devia- 

 ~' tion of this value from that gi\en by our 

 + - ° formula is only 1 in 1000. This accordance 



with Amagat's observations may be called excellent, especially when 

 we consider, that the calculation is as a matter of fact of the nature 

 of an extrapolation, albeit one which exceeds the limits of the 

 observation by very little only, as in Amagat's work the isothermal 

 of 0° C. is the only one which goes down to 100 atmospheres. We 

 can therefore now set aside the supposition, made before, that 

 Amagat's value al 100 atmospheres might be too high by 1 in 500. 

 This supposition was based on Schalkwuk's determinations up to 



